Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: Optimum ISO?  (Read 6736 times)
Jeff Welker
Newbie
*
Offline Offline

Posts: 18


« on: November 06, 2010, 05:40:04 PM »
ReplyReply

I'm new to the LL forum and returning to serious photography after a long hiatus. While I still have stains on my hands from working in my father's professional darkroom 30+ years ago, I am working hard at getting up to speed on the digital process. I got a copy of D-65's Lightroom Workbook today and have been reading with earnest. Seth/Jame make the statement that "Every digital camera has an optimum ISO setting"; and that "The best capture quality will be obtained using that [optimum] ISO". I've got a 5D MKII on order and would appreciate learning the best method to determine the optimum ISO for my camera. Sorry for the basic question; however, I want to get this right.

Thanks.
Logged
Policar
Jr. Member
**
Offline Offline

Posts: 56


« Reply #1 on: November 06, 2010, 06:25:19 PM »
ReplyReply

In general, the "optimum ISO" is the native ISO of the sensor, without gain (which increases noise in exchange for sensitivity), or "digital pulling" (which decreases highlight detail in the process of decreasing sensitivity but also reduced noise).

I believe the 5D's native ISO is 100, not 100% sure.

That doesn't mean it's always best to shoot at 100 ISO.  Some cameras have "highlight tone priority" or something to increase highlight detail by shooting a stop slower and pushing up dark areas digitally--and that requires a base ISO twice the camera's native ISO (200ISO on the 5D, I think).

Just think about digital ISO as you would with different films':  lower is cleaner and more detailed, faster means more flexibility in terms of stopping down or using a faster shutter.  In terms of absolute image quality, the "native" ISO is best; in terms of superior images, whatever allows you to get the f-stop/shutter speed combination you need while not looking too noisey is best.  And you can shoot up to 3200ISO and still get a great image, though you will notice noise.  The 5DII is so good you don't have to worry about shooting everything at a low ISO, as you would with a point and shoot.

Play around with the camera and see what you find acceptable.  Despite tons of obfuscating menus and stuff, digital cameras aren't substantially different from film slrs.  You may have to adjust your exposure technique over time, though.  That said, the image quality should trounce 135.  And sick camera, btw.  If I had the money that would be my top choice.
« Last Edit: November 07, 2010, 11:48:45 AM by Policar » Logged
01af
Sr. Member
****
Offline Offline

Posts: 294


« Reply #2 on: November 07, 2010, 05:01:17 AM »
ReplyReply

Most digital cameras don't have an optimum ISO setting. Instead, they have two.

At one ISO setting—usually the very lowest, typically ISO 50/18°, ISO 80/20°, or ISO 100/21°, in some cameras called 'pull'—the digital noise is at its optimum. i. e. lowest. At some other ISO setting—usually one or two stops above the lowest, typically ISO 160/23° or ISO 200/24°, in some cameras the lowest that's not marked 'pull'—the dynamic range is at its optimum, i. e. widest. These two optima usually don't coincide. Exceptions to this rule may exist, in particular at very large or very small sensor formats.

Most photographers go for the dynamic-range optimum whenever possible. Here, noise is so low that having even less noise is not considered worthwhile, especially not when it comes at the price of slightly reduced dynamic range. So the 'pull' settings usually get used only as a ND filter substitute rather than for their extra-low noise, when light is too bright for the desired aperture/speed combination at higher ISO settings.
« Last Edit: November 07, 2010, 05:03:17 AM by 01af » Logged
vandevanterSH
Sr. Member
****
Offline Offline

Posts: 626


« Reply #3 on: November 07, 2010, 07:29:37 AM »
ReplyReply

would appreciate learning the best method to determine the optimum ISO for my camera. Sorry for the basic question; however, I want to get this right.
**********
From the perspective of a hobbyist, I think that you are going to far into the "weeds" with "best method to determine the optimum ISO for my camera".  What is the optimum f/stop?, shutter speed, focal length? ISO?...just another variable to think about to obtain the best image for a particular situation.  Enjoy shooting RAW with your new 5D MKII and have fun post processing in LR3.

Steve
Logged
NikoJorj
Sr. Member
****
Offline Offline

Posts: 1063


WWW
« Reply #4 on: November 07, 2010, 08:30:24 AM »
ReplyReply

At one ISO setting—usually the very lowest, typically ISO 50/18°, ISO 80/20°, or ISO 100/21°, in some cameras called 'pull'—the digital noise is at its optimum. i. e. lowest. A
My 2c's : this one is not that useful, as these modes usually work only like overexposing at optimum ISO ; so, by exposing to the right at that ISO where dynamic range is highest, the same lowest level of noise is achieved.

Bottom line : to determine optimum ISO, take the maximum of the dynamic range plot of your camera in DxOmark.com.
Logged

Nicolas from Grenoble
A small gallery
Jeff Welker
Newbie
*
Offline Offline

Posts: 18


« Reply #5 on: November 07, 2010, 09:50:49 AM »
ReplyReply

From the perspective of a hobbyist, I think that you are going to far into the "weeds" with "best method to determine the optimum ISO for my camera".  What is the optimum f/stop?, shutter speed, focal length? ISO?...just another variable to think about to obtain the best image for a particular situation.  Enjoy shooting RAW with your new 5D MKII and have fun post processing in LR3.

Steve
Steve:

Thank you for the comments. I apologize for not being more detailed in my question - I will be shooting professionally. I am a private consultant to commercial land developers and they have been requesting that I provide them with photographic services (i.e. construction progress, archtectural, special inspections, etc). They are aware of my history shooting professionally for my father and want me to take care of their needs. Accordingly, I need to get my skills/knowledge updated from my film experiene to digital. This includes adopting best practices for my equipment.

Thanks again.
Logged
Jeff Welker
Newbie
*
Offline Offline

Posts: 18


« Reply #6 on: November 07, 2010, 09:56:16 AM »
ReplyReply

My 2c's : this one is not that useful, as these modes usually work only like overexposing at optimum ISO ; so, by exposing to the right at that ISO where dynamic range is highest, the same lowest level of noise is achieved.

Bottom line : to determine optimum ISO, take the maximum of the dynamic range plot of your camera in DxOmark.com.

Nick - I sincerely appreciate the comments. I went to the DxOMark site and (if I read the data correctly) it appears that 100 ISO is optimum. I want to make sure I understand your comments about exposing to the right at optimum ISO. Do you mean that with my 5D MkII I can get the highest dynamic range and lowest noise levels by exposing to the right at ISO 100? Sorry to be dense on this matter; however, I want to make sure I understand.

Thanks.
Logged
BartvanderWolf
Sr. Member
****
Online Online

Posts: 3461


« Reply #7 on: November 07, 2010, 10:09:19 AM »
ReplyReply

Do you mean that with my 5D MkII I can get the highest dynamic range and lowest noise levels by exposing to the right at ISO 100?

That's correct. However, to expose to the right, you need to know that the camera histogram is based on the JPEG thumbnail that's embedded in the Raw fle. That JPEG will not show the real headroom of the Raw data because a tone curve was applied to the JPEG. You'll get a feeling for how it's different when you pay attention during Raw conversion. Just shoot some bracketed exposures and note the JPEG histogram and compare with the Raw histogram.

As far as dynamic range goes, in difficult situations you can also use bracketing and later do exposure fusion or HDR tonemapping. But for now, first concentrate on getting the exposure of a single frame right, the rest will follow from that.

Cheers,
Bart
Logged
langier
Sr. Member
****
Offline Offline

Posts: 617



WWW
« Reply #8 on: November 07, 2010, 10:33:31 AM »
ReplyReply

Rather than finding absolute numbers, go out and get some shooting time with different iso settings under different lighting conditions. Get a feeling for what works best for you, rather than the numbers. Learn what your camera is capable of producing under different speeds and lighting and you'll be able to master this part of the craft.

You'll learn more than you can from running software generating a bunch of numbers and get more hands-on with your camera, thus get to know it better.

To get a reasonable approximation of what the chip sees, you've got to turn the camera's rendering intent to sRGB, turn down the contrast of the camera display a notch and run it at the normal color intent, not vivid, to get a closer version of what the chip sees. The image will appear flat.

The reason behind this is that your preview image is a rendered jpeg file created by the camera of the raw image. Those settings will skew the preview and thus give you even less accurate version of your final raw file. Since the file you are shooting is raw, color space is irrelevant.

Once in your raw workflow, you can then optimize your image, including setting the master file color space.

Nobody will ever see the numbers behind the image, it's the crafting of the final image for web, print, or otherwise that actually matters!
Logged

Larry Angier
ASMP, NAPP, ACT, and many more!

Webmaster, RANGE magazine
Editor emeritus, NorCal Quarterly

web--http://www.angier-fox.photoshelter.com
facebook--larry.angier
twitter--#larryangier
google+LarryAngier
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #9 on: November 07, 2010, 11:11:59 AM »
ReplyReply

Rather than finding absolute numbers, go out and get some shooting time with different iso settings under different lighting conditions. Get a feeling for what works best for you, rather than the numbers. Learn what your camera is capable of producing under different speeds and lighting and you'll be able to master this part of the craft.

I think that advise is the most sensible for those of us that can’t control every aspect of the light and exposure for all captures. While ISO 100 may indeed be the “optimal” for this or that sensor (best quality quality?), if shooting this way produces an unsharp image due to camera shake but shooting at ISO 400 doesn’t, with a bit more noise, seems clear what approach makes more sense. Sure, you have all the light you need to capture at the shutter/F-stop at ISO 100, why not. But I have plies of images I am happy with shot at ISO 1600 because that’s the only way I could get a capture in the first place. So yes, learn what the camera and raw converter is capable of producing.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Policar
Jr. Member
**
Offline Offline

Posts: 56


« Reply #10 on: November 07, 2010, 11:58:42 AM »
ReplyReply

Yeah, unless you're using a tripod and so shutter speed isn't a factor don't worry about "optimum ISO," just worry about what's good enough at a given print size.  That camera book offers that information under the assumption you have a cheaper (point and shoot) camera, with a smaller sensor, fewer pixels, and more noise.

You will be amazed how good the 5D is; up to at least 800ISO it will be clean enough that noise isn't an issue for prints up to 16X20 is my guess, especially with a little noise reduction in light room.  I recently visited with a semi-famous professional who needs fast shutter speeds above all else for his variety of work, and he'll go up to 3200ISO if needed with the intention of some images then being wall-sized prints.

If you're doing architectural photography and want the absolute best results, look into a tilt/shift lens, which should be very familiar to you if you used to shoot and process large format.  In which case you would want to use a tripod, in which case shutter speed is a non-issue and so you could use 100ISO under virtually any light.  HDR and exposure masking are also really popular with digital architectural photography, so that's another reason to use a tripod and bracket exposures.

But I think you'll be very surprised how good the 5D is across a wide range of sensitivities, especially when considering the printed image or final web output (not the camera original blown up to 100%, where noise is more visible).  That book was written with cheaper, less flexible cameras in mind.  The same way every lens has one "best" f-stop but it's not the only one you'd use, every camera's "best" ISO is good to know but nothing to get so concerned about using all the time.
« Last Edit: November 07, 2010, 12:04:27 PM by Policar » Logged
Jeff Welker
Newbie
*
Offline Offline

Posts: 18


« Reply #11 on: November 07, 2010, 12:08:46 PM »
ReplyReply

That's correct. However, to expose to the right, you need to know that the camera histogram is based on the JPEG thumbnail that's embedded in the Raw fle. That JPEG will not show the real headroom of the Raw data because a tone curve was applied to the JPEG. You'll get a feeling for how it's different when you pay attention during Raw conversion. Just shoot some bracketed exposures and note the JPEG histogram and compare with the Raw histogram.

As far as dynamic range goes, in difficult situations you can also use bracketing and later do exposure fusion or HDR tonemapping. But for now, first concentrate on getting the exposure of a single frame right, the rest will follow from that.

Cheers,
Bart

Bart - I've recently read Michael's "Expose (to the) Right" tutorial and similar comments by Jeff Schewe in "Camera RAW in Adobe Photoshop CS5" and his "From Camera To Print - Fine Art Printing Tutorial" video tutorial with Michael. As you've said, I need to play/familiarize myself with my 5D MkII to know how the LCD histogram relates to the actual RAW histogram. A lot of this stuff is the exact opposite of my film experience - very interesting but backwards.

Thanks.
Logged
Jeff Welker
Newbie
*
Offline Offline

Posts: 18


« Reply #12 on: November 07, 2010, 12:20:51 PM »
ReplyReply

Rather than finding absolute numbers, go out and get some shooting time with different iso settings under different lighting conditions. Get a feeling for what works best for you, rather than the numbers. Learn what your camera is capable of producing under different speeds and lighting and you'll be able to master this part of the craft.

You'll learn more than you can from running software generating a bunch of numbers and get more hands-on with your camera, thus get to know it better.

To get a reasonable approximation of what the chip sees, you've got to turn the camera's rendering intent to sRGB, turn down the contrast of the camera display a notch and run it at the normal color intent, not vivid, to get a closer version of what the chip sees. The image will appear flat.

The reason behind this is that your preview image is a rendered jpeg file created by the camera of the raw image. Those settings will skew the preview and thus give you even less accurate version of your final raw file. Since the file you are shooting is raw, color space is irrelevant.

Once in your raw workflow, you can then optimize your image, including setting the master file color space.

Nobody will ever see the numbers behind the image, it's the crafting of the final image for web, print, or otherwise that actually matters!

Larry:

Your comments are spot on - thanks. I can't wait to get some time with this camera. Long term I hope to find multiple "optimum" settings that work best in the shooting circumstances I may find myself in. I am simply looking for a good starting point to do some testing on my tripod and then work into other scenarios from there. While I can't control everthing, or sometimes very little, I want to end up with the best captures possible so that my post processing work is minimized - as much as possible.
Logged
Jeff Welker
Newbie
*
Offline Offline

Posts: 18


« Reply #13 on: November 07, 2010, 12:24:58 PM »
ReplyReply

If you're doing architectural photography and want the absolute best results, look into a tilt/shift lens, which should be very familiar to you if you used to shoot and process large format.  In which case you would want to use a tripod, in which case shutter speed is a non-issue and so you could use 100ISO under virtually any light.  HDR and exposure masking are also really popular with digital architectural photography, so that's another reason to use a tripod and bracket exposures.

But I think you'll be very surprised how good the 5D is across a wide range of sensitivities, especially when considering the printed image or final web output (not the camera original blown up to 100%, where noise is more visible).  That book was written with cheaper, less flexible cameras in mind.  The same way every lens has one "best" f-stop but it's not the only one you'd use, every camera's "best" ISO is good to know but nothing to get so concerned about using all the time.

I've also got the Canon TS-E 24mm f/3.5L II Tilt-Shift coming with the 5D MkII. I'm very excited to use these tools.
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7255


WWW
« Reply #14 on: November 07, 2010, 02:27:43 PM »
ReplyReply

Hi,

My view may be that it is minimum ISO and expose to the right. That is probably optimum. On some cameras it may be possible to raise ISO quite a bit without loosing dynamic range (spelled as DR, pronounced like "shadow detail"). Shooting at 100 or 200 ISO may be safest bet. Always expose to the right, that is make sure that non specular highlights are close to clipping but still not clipped.

Look at the example based on DxO data. The Canon camera gives about the same DR up to ISO 640. The Sony has best DR at 119 ISO. So for Sony the best ISO would be around 120 and for Canon it may matter little, up to 640.

Best regards
Erik



Best regards
Erik Kaffehr


I'm new to the LL forum and returning to serious photography after a long hiatus. While I still have stains on my hands from working in my father's professional darkroom 30+ years ago, I am working hard at getting up to speed on the digital process. I got a copy of D-65's Lightroom Workbook today and have been reading with earnest. Seth/Jame make the statement that "Every digital camera has an optimum ISO setting"; and that "The best capture quality will be obtained using that [optimum] ISO". I've got a 5D MKII on order and would appreciate learning the best method to determine the optimum ISO for my camera. Sorry for the basic question; however, I want to get this right.

Thanks.
Logged

Jeff Welker
Newbie
*
Offline Offline

Posts: 18


« Reply #15 on: November 07, 2010, 03:09:38 PM »
ReplyReply

Hi,

My view may be that it is minimum ISO and expose to the right. That is probably optimum. On some cameras it may be possible to raise ISO quite a bit without loosing dynamic range (spelled as DR, pronounced like "shadow detail"). Shooting at 100 or 200 ISO may be safest bet. Always expose to the right, that is make sure that non specular highlights are close to clipping but still not clipped.

Look at the example based on DxO data. The Canon camera gives about the same DR up to ISO 640. The Sony has best DR at 119 ISO. So for Sony the best ISO would be around 120 and for Canon it may matter little, up to 640.

Best regards
Erik

Erik:

Thanks much for helping me translate the DxO graph. Your comments are helpful.

Take care.
Logged
David Sutton
Sr. Member
****
Offline Offline

Posts: 874


WWW
« Reply #16 on: November 08, 2010, 01:06:15 AM »
ReplyReply

Hi Jeff. A couple of random thoughts.
Perhaps because the sensor elements are lined up in rows, instead of having sort of varying sized splodges in the emulsion, I find sharpness more an issue than iso with digital. I shoot at iso 100 unless the shutter speed is around say 1/20th to 1 second, when on my camera (5D2) on some surfaces, even with mirror lock up, just the shutter slamming open and shut can blur the image. Your tripod and the surfaces it stands on may give different results.
I start to really lose detail when stopping down to f16 and beyond due to diffraction.
Last night I started looking at the high iso abilities of the 5D2. After running the file though noise reduction there was not as much difference between the iso 100 and 6400 files in the mid to ¾ tones as I would have expected. Um, I'm going to have to reshoot the target to include complete blacks to better reflect a real world situation.
As the others here have said, there is no substitute for getting out in the real world and working out how to get the best from this camera. Have fun.
Logged

MrSmith
Sr. Member
****
Offline Offline

Posts: 815



WWW
« Reply #17 on: November 08, 2010, 03:48:04 AM »
ReplyReply

i only use 160/320/640 etc on my MkII. and never use 100 you get pattern noise at 100 but not at 160.   that's why those dxo charts are not the whole story as they never take into account the native iso.

Logged
MrSmith
Sr. Member
****
Offline Offline

Posts: 815



WWW
« Reply #18 on: November 08, 2010, 03:52:43 AM »
ReplyReply

i only skim read this thread before posting the above.
i disagree with a lot of what has been said, supposition and no user derived facts.

Fact 1. 160 is better then 100. no pattern noise in the blacks and the minimum of noise.
Fact 2. a good file is a good file and makes your job easier and you spend less time in front of the computer. if using 160 not 100 makes your life easier this is a good thing.
Logged
hjulenissen
Sr. Member
****
Offline Offline

Posts: 1666


« Reply #19 on: November 08, 2010, 04:06:18 AM »
ReplyReply

I shoot at iso 100 unless the shutter speed is around say 1/20th to 1 second, when on my camera (5D2) on some surfaces, even with mirror lock up, just the shutter slamming open and shut can blur the image. Your tripod and the surfaces it stands on may give different results.
Did you experiment with livemode in these cases? I was under the impression that livemode actually was slightly superior to MLU in terms of shake in recent Canon DSLRs.

-h
Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad